The Quiet Revolution: Operational AI That No One Notices (But Everyone Feels)

Most real change in healthcare does not occur with the announcing of a new policy. Rather, change occurs in the day-to-day workings of a complex system. It can be seen with the scheduling of shifts that occur without administrator input and operating rooms that start to run on time without the need for anyone to remind the staff of the new schedule. This is the kind of change that can be felt, even when the infrastructure of the system is not yet complete, the change that is felt when AI moves from aspiration to reality.

Right now, the focus of the innovation is on clinical AI, and the use of AI for diagnosis, generative note-taking, and predictive analytics are all front and center. However, much of the AI that is doing the heavy lifting is going largely unnoticed in the operational AI space. It is not about diagnosis. Rather, it is about the orchestration of the day to day activities of the staff and ensuring that friction is minimized to allow the day to run smoothly.

The Value That Does Not Announce Itself

Traditionally, the greatest value that AI has brought to the clinical space is in the moment of the clinical decision. This is the moment when the AI tells the clinician something they did not already know. However, the greater value is often below the surface of the day to day clinical activities. This is especially the case with operational AI. The value of operational AI is often seen with the prediction and management of bed turnover and the forecasting of staffing needs. The value of this type of operational AI is that it greatly alters the way the system operates, even if the change occurs without much fanfare.

The advantage extends beyond optimal operational hours for staff. It means no more waiting in line. When logjams dissipate on their own, the clinician and patient can experience an exception to the rule of healthcare distress, that being an absence of pressure. What would have been a frantic day, is only busy.

When Work is Done Behind the Scenes

If you distinguish the details, it becomes more apparent. It is the system that employs predictive analytics to determine the status of patient beds and prevents hallway waiting. It is the scheduling system that manages and fulfills task gaps to prevent staff shortages. It is the logistics system that ensures that all room prep is adequate and on time, as opposed to delayed and out of sync. None of these systems substitute for a clinician. Each of these systems contributes to reducing the cumulative burden of a full day across a whole building.

Operations That Are Seamless

The analogy of a nervous system is not too far afield. Operational AI does just that. It recognizes the points of building pressure and redistributes the burden, thus maintaining equilibrium, and allows the system to operate below the thresholds of noticeable discomfort. It smooths out operations without the intrusion of unnecessary dashboards or alterations to established workflows.

Consider mature predictive discharge modeling. A unit that has this will create the transport and paperwork several hours in advance to minimize wait time for the transport and to ensure the unit does not get surprised by a last-minute transport. The staffing holds together, and the turnover stays smooth because the day was forecast rather than discovered. There are fewer last-minute stressful situations and the staff on the unit would not be able to tell you why, because they are more focused on the task at hand. That is the goal with this kind of AI.

Maturity Looks Like Calm

The difference between most organizations and calm ones that have truly absorbed AI is that calm ones do not call it an innovation. They call it an infrastructure. They build an environment that runs predictably, the way a building runs when the lights simply work.

That is a milestone, and it is cultural more than technical. Absorbing AI means it becomes part of the normal functioning of the unit, the way the lighting is just part of how the unit runs.

When Work Equates to Non-Work

The people who implement these programs feel the strain of the contradiction. They must harness AI to bring a competitive edge that is demonstrable and visible to the board. But AI that is working is the most invisible. This dichotomy is the biggest challenge. Perhaps the best AI working in hospitals is the one that no one knows is there, the most invisible.

To really evaluate the success of invisible work, you cannot count only the visible work. Doing that rewards what gets noticed and overlooks the crisis that quietly never happened. Unseen value is of real worth. Sadly, it is not easily exemplified.

Being Invisible Does Not Mean You Are Not Being Watched

There is an important, but less desirable, side to the phenomena discussed in the previous section. A system that you do not consciously pay attention to is a system that you do not consciously monitor, and a drifting orchestration model will continue to generate suboptimal schedules and forecasts. They will be confident and erroneous, and no one will complain that anything is broken. A degraded operational AI hides the same way a healthy one does, behind the appearance that everything is running fine. The calm that such a model generates is fake. The real challenge is to keep a person watching the system precisely because no one else is, and to actually test if the system is performing its intended purpose, instead of assuming that the absence of indicators is a sign of success.

What Healthcare Actually Needs

Healthcare certainly does not need any more disruptive technology. It has had more than its fair share of that. Healthcare needs technology that improves flow and then completely removes itself from the system. When AI optimizes the flow of work, those within the system feel it without the need for active thought. The clinicians experience it as a day that holds together. The patients experience it as care that is delivered as expected.

The future healthcare professionals envision for AI is full of drama; an AI system that can diagnose illnesses with greater accuracy than the best human doctors. A quieter and more practical type of future is arriving. It does not happen all at once like an industry breakthrough. It resembles a system that has been running efficiently, with the steady absence of the factor that has been hindering it.

Christopher Hutchins Founder and CEO, Hutchins Data Strategy Consultants